This prototype demonstrates an early version of the Chat Assistant based on Retrieval‑Augmented Generation (RAG) over a scientific knowledge graph and the emerging SciGraphChat framework. It is a minimal viable product (MVP) currently undergoing development and is intended to showcase the core interaction concepts.

The demo provides two main user interface components:

•  Chat Panel (Text View): Enables conversational interaction with the system. Users can enter queries, ask questions, and receive generated responses grounded in retrieved scientific metadata. For each chat response, users can view the sub-graph, find related research, follow the links to the research artifacts (i.e., papers and supplementary materials), share the chat link with others, and export (Markdown, HTML, PDF) the generated answers and retrieved information.
•  Graph Panel (Graphical View): Offers a visual exploration environment in the knowledge graph including a graph search. A visual exploration helps users to understand complex and semantic relationships more easily by presenting the underlying knowledge graph in an intuitive and interactive way.

This demo is connected to a curated subset of the Open Research Knowledge Graph (ORKG) focused on software architecture research based on the dataset of the paper "Evaluation methods and replicability of software architecture research objects" published at the IEEE 19th International Conference on Software Architecture (ICSA).

Users can inspect relationships, navigate connected research artifacts, and gain insights through graph‑based representations. This demonstration focuses on illustrating how text‑based retrieval and graph‑based exploration can work together to support research‑oriented question answering. Additional features and extended functionality will follow in subsequent development stages.